Optimization of image quality and dose in adult and pediatric chest radiography via Monte Carlo simulation and experimental methods

Optimization in radiography aims to achieve adequate image quality to enable diagnosis by adjusting the dose to the patient. The figure of merit (FOM) is a well-known optimization method that balances image quality (IQ) and Dose. However, different authors use distinct definitions. This work aims to compare several FOM definitions to determine the most adequate for chest radiography optimization via Monte Carlo (MC) simulation. Moreover, this study aims to perform a systematic study, evaluating the influence of tube potential, additional filtration, detector technologies, antiscatter grids, and phantom thickness on FOM. The dose was evaluated by the mean absorbed dose (MAD), the entrance skin dose (ESD), and the effective dose (E). The signal-to-noise ratio (SNR) and the contrast-to-noise ratio (CNR) were the IQ indexes. FOM was calculated as IQ2/Dose. Two types of antiscatter grids were simulated. One grid had cotton fiber as an interspace material, and the other grid had aluminum. The simulated detectors were composed of CsI and a-Se, used in digital radiography (DR), and BaFBr, used in computed radiography (CR). The results showed that tube potentials close to 120 kV optimized the SNR for all cases of phantom thickness, dosimetric quantities, and additional filtrations. On the other hand, the op-timum tube potentials for CNR are dependent on these parameters. It was shown that optimizing the CNR provides better overall image quality. Additionally, MAD and E are suitable for optimization studies based on FOM. Additional filtration showed a better dose reduction, with 1 mm Cu having the best performance. Considering both optimization and clinical applicability, additional filters of 0.2 mm Cu + 2 mm Al are indicated. The optimum tube potentials are similar for all antiscatter girds evaluated. When different detectors are applied, the one with a higher effective atomic number provides larger FOM values and optimum tube potential. Experimental validation was performed only with the CR detector due to limitations in the detection technologies available. The dependency of FOM on the tube potential was slightly different between the experimental and MC simulation cases; however, there was an intersection between the optimum tube potential range for both cases. The results in this study yielded a general overview of optimization, analyzing several parameters, providing an improvement in other optimization studies that were limited by either one FOM definition or experimental factors such as one detector and/or one antiscatter grid configuration. (AU)